quiz bu lëkkale ak tegtal · Diggante Niveau

Reflexion ak laaj ak tontu ngir saafara sa bopp

Xoolaat boucle de feedback verbal bu Reflexion, jàngat, mémoire episodique, yaatuwaayu référence, ak risku feedback bu wóorul. Saytu firnde yi ak àpp yi.

Yooni njiit yu jëm ciReflexion ak Agent yuy jubbanti seen bopp
Laaj 1 de 8

Lan mooy màndarga biy wane ni agent Reflexion di 'jàngee'?

Xoolaatal laaj bu nekk8 laaj

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

  1. Lan mooy màndarga biy wane ni agent Reflexion di 'jàngee'?
    • Dafay denc xalaat yuñ bind ci muy tekki te du soppi diisaayu model bi
    • Dafay yeesal diisaayu model bi jaaraleko ci ajustement
    • Dafay yebbi xeetu model bu bees ginaaw bépp njuumte
    • Dafay laaj nit mu soppi etiketu done yi mu tàggat
  2. Ci kaadar Reflexion, lan la Evaluateur bi di def?
    • Dafay defar jëf ji ci topp ci ndawu liggéey bi
    • Dafay joxe poñ ci njariñu jéema defar siñaal feedback
    • Dafay efaase ay mémoire yu yàgg ngir sakkanal palaas
    • Dafay tekki liggéey bi ci beneen làkk
  3. Lan moo waral kalite siñaalu jàngat bi am solo lool ci Reflexion?
    • Ndax mooy wane ni model bi di jàngee
    • Ndax mooy wane ñaata GPU lañuy jëfandikoo
    • Ndax siñaal bu jaarul yoon mën na tax ndawu liggéey bi xalaat ci njàngale yu jaarul yoon
    • Ndax mooy tëral tàngooru xeetu làkk wi
  4. On which coding benchmark did the Reflexion paper report a 91% pass@1 result?
    • Nataal Net
    • COCO detection
    • WMT translation
    • Nit Eval
  5. Lan mooy risk bu dëggu bi ubbeeku ci agent yuy jubbanti seen bopp?
    • Mën nañu def loop budul jeex ngir 'fix' génnekaay bi jota jub
    • Dañu soxla done yuñ etiketee ci token bu nekk
    • Mënuñu dox ci hardware buñ miin
    • Dañuy yàq modelu base bi ba fàww
  6. Where does Reflexion keep feedback for a later attempt?
    • In updated model weights
    • In an episodic memory buffer supplied as text context
    • In a hardware cache shared across users
    • In the original training corpus
  7. What is the role of the Actor in Reflexion?
    • It changes the base model weights
    • It authenticates citations automatically
    • It generates the actions or attempt for the task
    • It sets the benchmark score after the run
  8. What kinds of feedback can Reflexion use?
    • Only gradient updates from fine-tuning
    • Only human ratings in every task
    • Only token likelihood from pretraining
    • Signals such as scalar outcomes or language feedback from external or simulated sources